Seedance 2.5 Video — 30-Second Stories in a Single Pass
Live on ImagineToVideo since August 7, 2026. Generate up to 30 seconds with synchronized audio.
Seedance 2.5 is ByteDance Seed's new-generation video model — and you can generate with it right now on ImagineToVideo. Up to 30-second audio-video clips in one pass, lip-synced dialogue, reference-driven generation and video-to-video, running on the official FAL API. This page covers the verified specs and how to get the most out of them.
Official Video
Watch: Seedance 2.5, as ByteDance Released It
The official release video ByteDance published for Seedance 2.5 on July 31, 2026. Every clip in it was rendered by ByteDance — none of it was generated on ImagineToVideo, and none of it has been independently reproduced.
Seedance 2.5 on ImagineToVideo — Verified Specs
30-Second Audio-Video in One Pass — Live
Generate up to 30 seconds of video with synchronized audio in a single pass — a complete story in one take, no stitching. Verified on ImagineToVideo: pick any duration from 4 to 30 seconds at 480p or 720p.
Lip-Synced Dialogue
Put spoken lines in double quotes inside your prompt and Seedance 2.5 generates matching lip movements and voice. Multi-subject interactions, camera moves and emotional tone are handled from the same prompt.
Reference-Driven Generation
Steer output with up to 9 reference images and 9 reference videos per generation on ImagineToVideo — keep a character, product or style locked across the whole clip. Timestamp-level editing of ByteDance's official spec arrives as we expand the integration.
Seedance 2.0 Still Available
The transition is yours to pace: Seedance 2.0 and 2.0 Fast stay live alongside 2.5 — same picker, lower credit cost, up to 1080p. Choose per generation which fits the job.
What Creators Generate with Seedance 2.5
Complete Short Stories
A 30-second story in one take with synchronized audio — live on ImagineToVideo. No multi-clip stitching, no seams, no drift between shots.
Reference-Heavy Production
Lock a character, product or art style across the whole clip with up to 9 reference images and 9 reference videos per generation.
Targeted Edits Without Regenerating
Feed existing footage through video-to-video and describe the change — restyle, fix or extend, billed at a discounted reference rate.
Cinematic Camera Work
Spoken spots and dialogue scenes: put the lines in double quotes and get matching lip movement and voice in the same pass.
How to Generate with Seedance 2.5
From prompt to a 30-second clip with audio in four steps
Open the Generator
Go to the ImagineToVideo generator — Seedance 2.5 is the default model, sitting at the top of the model picker with a Latest badge. No separate account or API key needed.
Write the Story and the Sound
Describe the scene, camera and mood in one prompt, and put any spoken dialogue in double quotes for lip-synced voice. Seedance 2.5 generates picture and audio together in a single pass.
Pick Duration and Resolution
Choose 4 to 30 seconds at 480p or 720p. Draft short and cheap at 480p, then re-run the winning prompt at 720p — or go straight to a full 30-second single take.
Add References or Rework Footage
Upload up to 9 reference images or 9 reference videos to lock characters and style, or feed existing footage through video-to-video — reference-based runs are billed at a discounted rate.
Seedance 2.5 FAQ
What is Seedance 2.5?
Seedance 2.5 is ByteDance Seed's new-generation video creation model, officially released on July 31, 2026. It focuses on long-form storytelling, multimodal reference and editing: ByteDance states it generates 30-second audio-video clips in a single pass, accepts up to 30 images, 10 video clips and 10 audio clips as reference, and supports timestamp-level editing.
Is Seedance 2.5 available now?
Yes. Since August 7, 2026 Seedance 2.5 runs on ImagineToVideo via the official FAL API — text-to-video, image-to-video and reference/video-to-video. It is the default model in our generator, and it also remains available in ByteDance's own apps (Jimeng AI and Doubao).
How long can a Seedance 2.5 video be?
Up to 30 seconds of audio and video in a single pass — verified on ImagineToVideo, where you can pick any duration from 4 to 30 seconds. ByteDance's project page additionally notes a clip can be extended for longer narratives; frame rate is not officially specified.
What reference inputs does Seedance 2.5 accept?
ByteDance's official spec allows up to 30 images, 10 video clips and 10 audio clips per generation. On ImagineToVideo you can use up to 9 reference images and up to 9 reference videos (1.8–30 seconds each, 30 seconds combined). Still to come: the full 30-image budget from the official spec, and audio references — both on our roadmap.
What is new in Seedance 2.5 vs Seedance 2.0?
Three verified differences: single-pass duration (up to 30 seconds vs 4–15), lip-synced dialogue generated from quoted lines in the prompt, and a stronger reference engine. One trade-off: Seedance 2.5 currently outputs 480p and 720p, while Seedance 2.0 goes up to 1080p — both stay available in the picker.
Can I use Seedance 2.5 on ImagineToVideo?
Yes — it is live and priced in credits by duration and resolution. Seedance 2.5 runs start at 40 credits, so the $0.99 Starter Pack (30 credits) covers a Seedance 2.0 clip while 2.5 itself takes a plan.
Generate Your First Seedance 2.5 Video
Seedance 2.5 is live with credit-based generation: up to 30-second clips, synchronized audio and reference-driven control.